Imagine risk analysis manager or compliance officer who can discover easily relationships like this: Big Bucks Café out of Seattle controls My Local Café in NYC through an offshore company. Such discovery can be a game changer if My Local Café pretends to be an independent small enterprise, while recently Big Bucks experiences financial difficulties.

We will demonstrate how such relationships can be discovered by using RDF graph database and a combination of open data and commercial datasets. Here is data architecture:

• Load GeoNames and DBpedia in a triplestore together with the links between them. The result is a repository of about 1 billion facts about all popular entities from Wikipedia and has quality spatial data about all sorts of locations on Earth
• Load in the same repository a company database from one of the Financial Data Services vendors (Factset, D&B, TR, etc.). Link companies from this database to DBpedia by matching ticker symbols. Link locations from this database to Geonames
• Map the major classes and relationships from these datasets to FIBO
• Make sure the triplestore is configured to support inference over transitive properties, e.g. geo:parentFeature, which encodes sub-region relationships, and FIBO’s fibo-fnd-rel-rel:controls, for control of a legal entity
• Make a SPARQL query about businesses in a specific region or industry in USA that are controlled indirectly by other USA businesses through company registered in an offshore zone

Sounds like an idealized utopic scenario that would be impractical to implement? We demonstrate who this can be implemented with GraphDB within a month of work.